Skip to main content

Full text of "A generalized Lieb-Liniger model"

See other formats


arXiv: 1508.0201 lv4 [cond-mat.quant-gas] 19 Nov 2015 


A generalized Lieb-Liniger model 

Hagar Veksler and Shmuel Fishman 

Physics Department, Technion- Israel Institute of Technology, Haifa 32000, Israel 

Abstract 

In 1963, Lieb and Liniger solved exactly a one dimensional model of bosons interacting by a 
repulsive 5-potential and calculated the ground state in the thermodynamic limit. In the present 
work, we extend this model to a potential of three 5-functions, one of them is repulsive and the 
other two are attractive, modeling some aspects of the interaction between atoms, and present an 
approximate solution for a dilute gas. In this limit, for low energy states, the results are found to 
be reduced to the ones of an effective Lieb Liniger model with an effective 5-function of strength 
c e ff and the regime of stability is identified. This may shed light on some aspects of interacting 
bosons. 


1 



I. INTRODUCTION 


The physics of Bose gases is a fascinating and complicated field of research. Since it 
involves a many-body problem, analytical results are rare and in some parameter regimes, 
one can use approximations to describe experimental systems with very good accuracy. For 
example, for a weakly interacting Bose gas, Mean-Field approximation can be used to reduce 
the many body Hamiltonian into a one-body non-linear Schrodinger Equation, the Gross- 
Pitaevskii Equation OT- In the opposite limit, a strongly interacting one dimensional Bose 
gas can be mapped into a gas of free Fermions (Tonks-Girardeau gas, see, for example, 
M). Exact solutions in other regimes are highly desired. 

Simple models like the Lieb-Liniger (LL) model, that may not have direct experimental 
realization, may alert us to unexpected physical phenomena that are overlooked when “rea¬ 
sonable approximations” are made and motivate experiments [6H8]. The model introduced 
in the present work is of this type. 

In their seminal work from 1963 [S], Lieb and Liniger managed to solve exactly a one 
dimensional model for interacting bosons. They considered the Schrodinger equation for N 
particles interacting via a 5-function potential 


h 2 A a 2 


N 


E 


2 m ' dx? 

3=1 3 


'Y16{xj-x a ) 


S = 1 

j>s 


ip = Eip 


( 1 ) 


where Xj is the coordinate of the j-th particle and c is the amplitude of the 5 function. 
Making a Bethe ansatz [19] 

N 


^ Ou,..., X N ) cx (-!) [P1 exp < i ^ x n k Pn > J | 


n =1 


j>s 


imc 

k Pj - kp s - -^-sign l x 3 - x s) 


( 2 ) 


where kp n are the k vectors obtained by the permutation P (where [P] is its parity) of the set 
ki ,..., kj\r. Lieb and Liniger wrote Bethe ansatz equations for the fc’s by imposing periodic 
boundary conditions on a ring of length L [TO , TTj, 


N ,_9 /7 , \ • N _)_ imc 


rT i T-r & (kj - k h ) + imc T r 
exp {ikjL} = 11 ,, _ = 11 


h 2 (kj-k h ) 

h 2 (kj — kh) — imc 1 — , 

h^j y 3 h^j n 2 (kj-k h ) 

These N coupled equations are solved numerically and the energy 

N 


( 3 ) 


h 2 \ - 

E = — ) k 

9 m < ^ 


2 

2 m ^ 2 
3=1 


( 4 ) 


2 











is calculated for the ground state and the excitations mmm- 

In the present work, we study a simple model which takes into account the range of inter¬ 
particle interactions without giving up the mathematical simplicity. It is a generalization of 
the LL model [9] where in addition to the repulsion there is also attraction. It is defined by 
the Schrodinger equation for N interacting particles of mass m, 


2m ^ dx 2 
3 =i 3 


N 

+ c 0 ^5 (xj - x s ) 

s =1 
j>s 


N 

+ Q y, 8 (Xj - X, 

s =1 
j>s 


N 

l) + Cl ^2 S (' X j - X S +l) 

s=l 

j>s 




Ell) 


( 5 ) 

where the inter-particle interaction is modeled as a sum of three 5-functions: The central one 
is repulsive (co > 0) while the peripheral ones are attractive (q < 0). This model is inspired 
by the Van-der-Waals potential which has repulsive and attractive regimes. By adjusting 
the parameters Co,q,/ of (|5]), one can model scattering from many inter-particle potentials 


In section II, we present Bethe ansatz equations for two bosons interacting via three 5- 
functions interaction potential and in section III an approximation is introduced, that allows 
to extend the LL Bethe ansatz equations to an arbitrary number of particles. The ground 
state solution for the approximate equations is found in section IV. Section V specifies 
the parameters of the regime where the gas is stable. The results and their experimental 
relevance are discussed in section VI. 


II. BETHE ANSATZ EQUATIONS FOR TWO BOSONS INTERACTING VIA 
THREE 5-FUNCTIONS INTERACTION POTENTIAL 

We start by writing Bethe ansatz equations for a simple case where there are only two 
bosons. In this case, the equations are intuitive. 

Consider two bosons of mass m trapped on a ring of length L and interact according to 
(|5|. It is convenient to write the wave function ij) in terms of center of mass coordinate, 
rq = (xi + X 2 ) /2 and relative motion coordinate, r 2 = (aq — x 2 ) /2, 

(ri,r 2 ) = ~^=e lkiri (j) (r 2 ) (6) 

where k\ = 2ixn/L and n is an integer so that periodic boundary conditions are satisfied. At 
the center of mass frame of reference, k\ — 0 and the wavefunction of the relative motion, 


3 





</>(r* 2 ), satisfies the Schrodinger equation 

Yp cP 11 1 

-^-^2 + ^ c oH r 2 ) + ^ c iH r 2- l /2) + - Ci d(r 2 + l /2) <f) (r 2 ) = E<j> (r 2 ) (7) 

which can be written also as 
’ h 2 <9 2 

~2m ^r 2 + C °^ ^ + ° 1 ^ ^ r2 ~ ^ + Q< * ^ + ^ ^ ^ = 2Ei ^ ^ ' ( 8 ) 

As usual in such cases, the wave function takes a different functional form in each of the 
four intervals [— |], [—|,0], [0, |] and The result for 0 (r 2 ) is 


4> (r 2 ) = C cos ( k 2 r 2 ) + iC sign (r 2 ) Q 0 sin ( k 2 r 2 ) + 


Q/sign (r 2 - 1/2) sin k 2 (r 2 - 1/2 ) + (/sign (r 2 + Z/2) sin fc 2 (r 2 + Z/2) 


where C is a normalization constant while Q 0 and Qi should be determined. They are easily 
determined for the 5-function interaction since the jump of the derivative at the locations 


of the 5-function satisfies 


Ac/' (r*) = 


dcj) (r 2 ) 


2 lrJ+0+ 


dcj) (rg) 
dr 2 


2 Ir^+O - 


where r 2 = 0, c* = cq or r 2 = Z/2, c* = q. This results in two equations for Q 0 and Qi 


ik 2 Q 0 [k 2 j = jl + 2 iQi sin [k- 2 l/2 


leading to 


ik 2 Qi = 


Qi \k 2 ) — 


cos (k 2 l/2)j + i <5 0 sin (k 2 l/2)j + Qi sin (jc 2 l 


ik 2 h 2 mci cos (k 2 l/2)j + sin [k 2 l/ 2^ 
/c 2 h 4 — 2m 2 c 0 ci sin 2 (k 2 l/ 2 ) - k 2 h 2 mci sin (k 2 l 


k 2 should be determined to ensure periodic boundary conditions (/' (r 2 ) = (/' (r 2 + L/2). In 
addition, cf) (r 2 ) = cj> (—r 2 ) so that (/' (r 2 ) = —</' (—r 2 ). Therefore, in particular, </' (r 2 )| r ., =L / 4 
must vanish, leading to 


0 ik 2 Ll2 _ 


1 — <5o ( k 2 ) — 2Qi (/c 2 ) cos M/2 


1 + Qo ( ^2 ) + 2<5; (/c 2 ) COS k 2 l/2 


4 



Now, we return to coordinates x\,x 2 . For this purpose, we use the relations: k\ = 
ki + k 2 ^/2,k 2 = (ki -k^j/2,ri = (aq + x 2 ) /2 and r 2 = (xi — x 2 ) /2 resulting in 

hn + k 2 r 2 = k 1 x 1 + k 2 x 2 (15) 


and 


hn - k 2 r 2 = k 2 x i + kix 2 . 


(16) 


The function -0 of (J6]) takes the form 


0 (x u x 2 ) = C [e i(fcia:i+fc23;2) + e i{ - k2Xl+klX2) ] (17) 

+Csign (aq - x 2 ) Q 0 (h - k 2 ) ( e ^ klXl+k2X2) - e ^ k2Xl+klX2) ) 

+CQi (k x - k 2 ) sign (aq - x 2 - l) [ e i ^+ k ^ e ~i(ki-k 2 )i /2 _ e *(fa*i+fci* a ) e t(fci-fe 2 )i/ 2 j 
+CQi {h - k 2 ) sign (aq -x 2 + l) _ e i(fc 2 * 1 +fc 1 * a ) e -i(fc 1 -fc a )t/ 2 j _ 


The periodic boundary condition 0 (aq + y,aq) = 0 (aq — %,x 2 ) results in 


and 


c ik,L 1 - Qo (*'1 - k 2 ) - 2Q, (ki - k 2 ) 

i cos ((A)! — k 2 ) 1/2 ) 

(18) 

1 + Qo (Aq ~ Aq) + 2Qi (. Aq — k 2 ) 

cos ((Aq — k 2 ) 1/2) 

ik 2 L 1 + Qo (Aq — k 2 ) + 2 Qi (. Aq — k 2 ) 

cos ((Aq — k 2 ) 1/2) 

(19) 

1 - Qo (k\ - k 2 ) - 2Qi (/.:, - k 2 ) 

i cos ((Aq — k 2 ) 1/2) 


which are identical to (14) (under the assumption k\ = 0, namely, in the center of mass 
frame of reference). In the derivation we used the fact that aq — x 2 —> aq — x 2 + L involves 
rotation around the circle and consequently all the signs are changed. 


III. APPROXIMATE BETHE ANSATZ EQUATIONS FOR AN ARBITRARY 
NUMBER OF BOSONS 

The two particle solution cannot be simply generalized to an arbitrary number of particles 
since for small interparticle distances, 


|Xj — aq| < /, 


( 20 ) 


the sign function in equation corresponding to varies substantially. For small l, the effect 


of the regime (20) may be negligible as demonstrated in what follows. This is reasonable 


5 






for a dilute gas where l <C L/N. In such a situation, the LL solution is valid with the 
replacement ) — Q o + 2Qi cos ((kj — k s ) 1/2), leading to 


V’O&i ,-,x N ) = c 


p 


N 


exp 




( 21 ) 


77.—1 


]^[ {1 + (Q 0 (kj - k s ) + 2Qi (kj - k s ) cos ((kj - k s ) 1/2)) sign (xj - x s )} 


j>s 


and 


e ikjL _ 


n 


1 ~ Qo (kj - k s ) - 2Qi (kj - k s ) cos ((kj - k s ) 1/2) 
1 + Qo (kj - k s ) + 2 Qi (kj - k s ) cos ((kj - k s ) 1/2) 


( 22 ) 


The kj are distinct, namely, the wave function vanishes if kj = k s for s ^ j as was shown in 
the original work of LL [9]. 


In the region where inequalities (20) are not satisfied for any of the particle pairs, the sign 
functions in (|9]) are all equal. Therefore, in this regime, (21) is a solution with the spectrum 
(22). There is a Hamiltonian that is different from the original one, for which (21) and (22) 


are eigenfunctions and eigenvalues even if some of the inequalities (20) are satisfied. It is just 
defined by the eigenfunctions and eigenvalues. For 1 = 0, this Hamiltonian and the original 
one are identical. If the spectrum and the L 2 -norm of the eigenfunctions are continuous in 
l, the relative difference in the spectrum and the wavefunctions (in the L 2 -norm) goes to 
zero in the limit l —> 0. If they are also differentiable as a function of /, then the relative 
difference behaves as Nl/L. 

We show that for the low energy states, the 3-5 function system can be replaced by a 
system with one (5-function of strength c e ff- 

We assume 

(kj - k s ) l < 1 (23) 

for all wave vectors kj. In section V, we show that this limit is relevant for the ground state 
and low-lying excitations of a dilute gas since k max < (const) ^ is small. In the leading order 
in kjl, 

Q 0 + 2 Qi cos ((kj - k h ) 1/2) « 


m 


h 2 ko 


Co + 2 Ci + 


mcil 

~w~ 

[2c 0 + 2c, + mc p l + 

itiCqI 

2h 2 


-| m 2 coQ/ 2 mcil 

1 2 s 4 h 2 



(24) 


the error is of the order Nl/L. Comparing (22) with (|3]), one finds that for small kjl, 
the behavior of the present problem is indeed similar to the one found for one (5-function 


6 





















potential of strength 


c e // ~ c o + 2C, + 


mcil 

K 2 

[2c 0 + 2c, + + 

itiCqI 


1 ra 2 coC/Z 2 mcil 

1 2ft 4 ft 2 



(25) 


in the leading order in kjl. Eq. (25) is the main result of the present work, and it enables one 
to understand the physics of the three 5-functions interaction in terms of the one 5-function 
interaction. Of particular interest are situations where c e // is very different from Co + 2q (the 
total strength of interactions). In order to find such situations, we define the parameters 


r = ci /cq and x = mcol/h 2 and rewrite (25) as 

Ce f f 


= 1 + 


rx (2 + 2 r + rx + ^x) 
c 0 + 2q ' (1 + 2 r) (l — \rx 2 — rx) 
rx + 1 + 2r 


(26) 


(1 + 2r) (l — | rx 2 — rx) 
For weak interactions, (r<l and ra < 1 ), 


-eff 


Co + 2 ci 

However, for very strong interactions (x —> oo), 

°eff _ - 2 


1. 


(27) 


-+ O' 


(28) 


c 0 + 2 ci (1 + 2r) x 

This is a surprising result. It is instructive to analyze the behavior of c e /// (cq + 2c,), Eq. 


(26), as a function of x. We are interested in the regime x > 0 and —0.5 < r < 0. At x = 0, 


the derivative of (26) is negative and therefore the function decreases. At 

(1 + 2r) 


Xo = 


(29) 


it turns out that c e ff = 0 (even though c 0 + 2q ^ 0). Higher values of x result in negative 
values of c e ff, namely, the effective interaction is attractive (even though Co + 2 q > 0). 
Schematic description of c e ff/ (c 0 + 2q) is given in Fig. 1. 


7 





















Figure 1: (Color online) Schematic description of c e ff/ (co + 2 q) as a function of x for r = ci/cq = 
— 0.25. The inset expands the region where c e ff changes its sign and the red dot is (xq,0). 

The result c e ff = 0 at x = Xq is verified numerically (see Fig. 2) and will be discussed in 
what follows. In the two particle case it is exact. For a related result see [m nsj. 


IV. GROUND STATE ENERGY 


In the previous section, we derived the approximate Bethe ansatz equations (22) for N 


bosons interacting by a three 5-function potential (|5]). The solution for these N coupled 
equations, (ki, /c 2 , • • •, fcjv), can be used to calculate the energy of the gas 


N 




(30) 


3 = 1 


In the ground state, \kj\ are minimal (but yet kj are different, as in the original work of LL 

0 )- 

Lieb and Liniger [9j managed to calculate the ground state energy in the thermodynamic 


limit (N —y oo) by solving only two coupled integral equations (35) and (36) (instead of 


N equations of the form (22)). Here, we obtain similar equations by using the logarithmic 
G (kj) = kjL + 9 (kj — k s ) = 2n (n 


form of (22), 


N+l 




(31) 


where 


9 (k) = i In 


Qo ( k ) + 2 Qi ( k ) cos (kl/2) - 1 
_Qo ( k ) + 2 Qi ( k ) cos (kl/2) + 1 


(32) 



















We see that if l = 0, the ground state corresponds to the choice rij = j, (j = 1,... ,N). 
This is true also for l ^ 0, as long as 9 is a monotonic increasing function of k. To see this, 
assume kj > k rni then, by monotonicity of 0, 9 (. kj — k s ) >9 (k m — k s ) for all s, therefore 
G (kj) > G (k m ) and G (kj) is monotonic. Since 9 is an odd function, G (kj = 0) = 0. The 
kj for the ground state are the smallest possible in absolute value, hence, we choose rij = j 
for the ground state. Therefore, in the monotonic regime, 


L (kj + i — kj) + (kj .|_i — kj) 9' (kj — k s ) = 2tt 

sj- j 


(33) 


where 9' (k) = 89 (k)/dk and kj and k 3+ \ are adjacent wave numbers. Typically, 9 i 


is 


monotonic and (33) is justified at the regime where (23) holds (see Sec. V for more details). 


The density of states per unit length in k space, is defined as 

1 


P ( k j) = 


L(kj+i kj) 


and satisfies 


f A N 

/ dkp (k) — —. 
La L 


It is used to write (33) in the form 


P (k) ~ ~ / dqp (q) 9' (k - q) = —. 


2vr 


I-A 


2tt 


(34) 


(35) 


(36) 


Here, A is the Fermi momentum (this should not be confused with fermionic systems!). The 

/fn - “ / dkp (k) k 2 . (37) 


ground state energy (30) is 


2 m 


r -A 


In order to solve Eqs.(35) and (36), we change into dimensionless variables: 

cpnL 


k 


c 0 m 


Z A’ "° Ah 2 ’ 


&i — 


cim 

M 2 ’ 


7o = 


CqUiL 
h 2 N ’ 


7 1 = 


h 2 N ’ 


d = M. 


(38) 


and (37) are, respectively |9j, 


-l 


In these variables, p(z) = 9 (A z), the density of states is g (z) = p (A z), and Eqs. (35), (36) 

(39) 

(40) 


7o 


dzg (z) = a 0 , 

J-l 

If 1 , 1 

/ dyr} (y-z)g (y) = — 


and 


e = 


2vr j_ 1 

2mEoL 2 

h 2 N 3 


2vr 


7o ^ 


= -3 / d v9 ( y ) y 


a, 


(41) 


o J -1 


9 


















How does one solve Eqs. (39) and (40)? First, it is necessary to choose values for «o, on and 
d. These values are related to the parameters of the Hamiltonian via the Fermi momen¬ 
tum A which is unknown at this stage. One should only keep in mind that ai/a^ = Ci/cq 
and therefore the ratio ai/ao does reflect the ratio between attraction and repulsion in the 


Hamiltonian. The integral equation (40) (with parameters a^,ai and d) can be solved nu¬ 


merically. The solution, g (z), should be substituted in (39) in order to find 70 . By repeating 
the above scheme for different parameters, it is possible to plot the dimensionless energy e as 
a function of the dimensionless interaction strengths q 0 and 7 ; and the dimensionless length 
d. for small values of d, the energy e depends only on the effective strength of interaction, 


that is (25) in dimensionless units, 


leff = 7o + 2 7* + 


aid (270 + 27 1 + cgd'y 0 + la 0 d^ 0 ) 
1 — baoaid 2 — aid 


(42) 


The significance of this effective strength of interaction is demonstrated in figure 2. In this 


figure, we present the solutions e(a 0 ,cq,<f) that were calculated by solving (39), (40) and 


(41). In Fig. 2(a), the energy is plotted as a function of the total interaction strength 70 + 27 ; 
and different choices of d are represented by different colors. It is clear that the effect of d is 
not negligible. Even at the regime d ^ 1, it is evident that the value of l has a strong effect 
on the ground state energy. Furthermore, even for a given value of d, the total interaction 
strength 70 + 27 ; (which is proportional to Co + 2 q) is not in one to one correspondence with 
the energy and therefore cannot be used to characterize the gas. Fig 2(b) shows that in the 


regime d 1, the energy indeed depends only on 7 e ff of (42). The results are consistent 


with (26). 


10 













(a) 


(b) 




Figure 2: (Color online) The dimensionless energy e of (41) as a function of dimensionless interaction 
strengths for q = — cq/4 (namely, r = —0.25) and 0 < «o < 30. (a) e as a function of 70 + 2 7 /. 


Different lines represent different choices of d of (38), from top to bottom: d = 0 (blue), d = 0.02 
(green), d = 0.04 (red), d = 0.06 (turquoise), d = 0.08 (purple). Points where the effective 
interaction is attractive were excluded from the figure (these were supposed to appear in the bottom 


purple curve in the regime x > xq = 2, see Eq. (29)), so that the highest value of x which does 
appear in the figure is x = 1.94 and it corresponds to the purple point (7.7,0.114). (b) The energy 


e of (a), plotted as a function of 7 e // (Eq. (42)). 


V. REGIME OF STABILITY AND DEFINITION OF DILUTE GAS 


The Bethe ansatz equations (22) and the effective interaction (25), are valid only where 


(23) is satisfied. Therefore, it is important to identify the regime where (kj — k s ) l -C 1. 
In the original LL model, the ground state energy and the values of Us are maximal for 


strong interactions, c —>■ 00 , where k n = ‘^fn and n' s are integers n = — y,..., y. Then, the 


N 


N 


maximal absolute value of k is k max = ^ and for all j, s, 


(kj — k s )l < 


2t tNI 


(43) 


For dilute gas, the inter-particle separation L/N is much larger then the interaction range l 


and (23) is satisfied. 


The same argument can be written for the three 5-functions interaction potential (J5]) . 


11 














If 9 of (32) is a monotonic increasing function of k, the ground state is given by rij = j. 


(j = 1, ■■■ ,N) and k 

max N • 

Let us analyze the function 9 (k) and identify the regime of parameters where it is 
monotonic, first, note that 9 (k) is monotonically increasing if and only if / (k) = 
j [Qo + 2 Qi cos (kl/2)] is monotonically increasing. For k —> 0, / (k) = —p^c e // and there¬ 
fore it is monotonically increasing as long as c e // > 0. Hence, if c e // > 0, there exist some 
k* (which depends on the parameters c 0 , q , Z and does not depend on L and TV since 6 is 
independent of these variables) such that for all k < k*, 6 (k) is monotonically increasing. 
For the ground state, the states with the smallest \kj\ are occupied, namely, rij = j with 
j = 1,..., TV, and 

G (. kj) < nN. (44) 

9 is an angle variable and therefore it is bounded (actually, for very small k, 9 = —i r). Hence 

N 

\kj\ < (const)— (45) 

Jb 

and can be made arbitrary small. Now, by increasing L (or decreasing TV), one may tune 
the value of k max such that the conditions 

k <r k* 

™max ^ rv 


and 


kmaxl 1 


are satisfied simultaneously, the regime (23) of dilute gas is achieved and our solution is 
correct up to a term of order Nl/L. 

For a dilute gas there is a range of parameters where c e ff > 0 and the solution is stable. 
There is also a range of parameters where c e // < 0 and the system is unstable. 


VI. SUMMARY AND DISCUSSION 

In this paper, we analyzed a one dimensional dilute Bose gas for an extension of the LL 
model defined by (|5]). By dilute gas, we mean that / -C L/N, that is, the effective size 
of a particle l (for example, the Van-der-Waals radius of an atom) is much smaller than 
the inter-particle distance. Using this assumption and the Bethe ansatz, we derived the 


approximate equations for the spectrum (18) ,(19), (22). In principle, these can be solved 


12 







numerically. For low energies in this situation \kjl\ -C 1 and the model can be approximated 


by a LL model with one 5-function of strength c e ff given by (25) and in dimensionless units 


by (42). The error of this approximation is of order Nl/L. This is a good approximation 
for the dilute gas. The effective interaction c e // depends on q and Co but also on the ratios 
between the characteristic potential energy scales, q/Z and co/Z, and the kinetic energy scale, 
h 2 /ml 2 , of a particle trapped in a well of length Z. 

Naively one would expect that for small kj, c e ff ~ Cq + 2q. It turns out to be correct for 
relatively weak interaction energy. For stronger interactions, c e ff becomes very small and 
even changes its sign (see Fig. 1). Note that this result holds also for the two particle case 
where it is exact. It is a surprising result, verified numerically in Fig. 2 and its experimental 
verification should be considered a challenge. The knowledge of c e // enables to calculate 
the ground state and the low excited states if the conditions for stability are satisfied. In 
section IV, the ground state is calculated in the thermodynamic limit for a dilute gas. In 
particular, it is demonstrated to depend on all parameters via c e ff. We have shown that 
for a dilute gas there is a regime of parameters where c e ff > 0 and therefore the system is 
stable. For other parameters, c e // < 0 and the dilute gas is unstable. In this regime, the 
results of H2H8] regarding dynamics of attractive gas might be realized. If the gas is not 
dilute, we cannot determine the stability of the system. This theoretical model enables to 
predict qualitative features of interacting bosons for realistic systems. 

The potential (J5]) can be realized, for example, in optical lattices [Hj with tight harmonic 
trapping along two perpendicular directions (E <C hco±) and almost flat potential along the 
third direction. The inter-particle interactions are in three dimensions and can be modeled 
by a “delta shell” potential 


V(r) = 


3cph 

4 rf n mui± 

cih 

2r'£ ut e ou tmuj_ 

0 


for r < r in 
for r out < r < r out + £ ou t 
otherwise 


(46) 


Where r in ,£ out —> 0. In a previous work [16j, we calculated the scattering length a and 
the effective range r e of the potential ( J46| ) (see App.A of [IB]), wrote a three dimensional 
Schrodinger equation and integrated it over two axes to obtain a one dimensional equation 
of the form (|5]). This leads to the relations 


Tout = \/Sl/2, 


(47) 


13 








1 


a = 


(cq + 2 ci ) 


and 


r, = 


Ahw_ 

2cil 1 2 3 4 5 6 7 2 a 

a (cq + 2q) 3 


(48) 


(49) 


As seen from (25), for l = 0, c e ff is proportional to the scattering length. However, for 
l 7 ^ 0, c e ff cannot be expressed in terms of a and r e . Therefore, it motivates introducing an 
effective scattering length that dominates the spectrum. 

From an experimental point of view, it looks that c e // is the only quantity that one can 
measure in order to characterize the inter-particle interactions (because it determines the 
spectrum). Hence, it makes sense to define an effective scattering length 


a eff 


Ce f f 

Ahuj± 


(50) 


This scattering length, which includes corrections originating in the non-vanishing interac¬ 
tion range, is unique for one dimensional bosonic systems. 


Acknowledgments 

We thank Eliot Lieb and Avy Soffer for suspecting an error in the original version of the 
work and Daniel Podolsky, Yoav Sagi and Efrat Shimshoni for illuminating and informa¬ 
tive discussions. The work was supported in part by the Israel Science Foundation (ISF) 
grant number 1028/12, by the US-Israel Binational Science Foundation (BSF) grant number 
2010132 and by the Slilomo Kaplansky academic chair. 


[1] C. Pethick and H. Smith, Bose-Einstein Condensations in Dilute Gases (Cambridge University 
Press, 2002). 

[2] L. Pitaevskii and S. Stringari, Bose-Einstein Condensation (Oxford science publications, 2003). 

[3] F. Dalfovo, S. Giorgini, P. Pitaevskii, Lev, and S. Stringari, Rev.Mod.Phys 71, 463 (1999). 

[4] M. Girardeau, J. Math. Phys. 1, 516 (1960). 

[5] M. Olshanii, Phys. Rev. Lett. 81, 938 (1998). 

[6] B. Paredes et ai, Nature 429, 277 (2004). 

[7] T. Kinoshita, T. Wenger, and D. S. Weiss, Science 305, 1125 (2004). 


14 







[8] F. Meinert et al., Phys. Rev. Lett. 115, 085301 (2015). 

[9] E. H. Lieb and W. Liniger, Phys. Rev. 130, 1605 (1963). 

[10] V. E. Korepin, N. M. Bogoliubov, and A. G. Izergin, Quantum Inverse Scattering Method 
(Cambridge University Press, 1997). 

[11] M. Takahashi, Thermodynamics of one-dimensional solvable model (Cambridge University 
Press, 2005). 

[12] E. H. Lieb, Phys. Rev. 130, 1616 (1963). 

[13] Z. Ristivojevic, Phys. Rev. Lett. 113, 015301 (2014). 

[14] T. Cheon and T. Shigehara, Phys. Lett. A. 243, 111 (1998). 

[15] T. Cheon and T. Shigehara, Phys. Rev. Lett. 82, 2536 (1999). 

[16] H. Veksler, S. Fishman, and W. Ketterle, Phys. Rev. A. 90, 023620 (2014). 

[17] P. Calabrese and J.-S. Caux, Phys. Rev. Lett. 98, 150403 (2007). 

[18] P. Calabrese and J.-S. Caux, J. Stat. Mech. , P08032 (2007). 

[19] For those who are not familiar with Lieb-Liniger solu¬ 

tion, we recommend on lecture notes by Mikhail Zvonarev, 
http://cmt.harvard.edu/demler/TEACHING/Physics284/LectureZvonarev. pdf and on 

the books |10[ [IT] . 


15